Shape Optimization of Acoustic Mufflers Using Radial Basis Function Networks Expression

Abstract

A Topology optimization attracts attention as a useful technique because of its flexibility in the design and applied to the various fields. Focusing on the acoustic field, there are some studies on the optimization design of an acoustic muffler by topology optimization. In the conventional topology optimization method, however, an optimal shape depends on the initial shape since sensitivity analysis is used. This paper presents a design method of an acoustic muffler using topology optimization based on genetic algorithm and shape expression method by Radial Basis Function Network. This method realizes obtaining a global optimal solution and representation of complex shapes with a few parameters. The proposed method is applied to the shape optimization problem of an acoustic muffler. The optimized result of an acoustic muffler is presented to show usefulness of the proposed method.